In the global AI race, can India run fast enough to catch up with leaders?

India has set its sights on becoming a global leader in artificial intelligence, with a strong push towards developing its own AI ecosystem. This ambition is evident in recent announcements from the Ministry of Electronics and Information Technology (MeitY). In March 2024, then-minister Rajeev Chandrasekhar announced a Rs 10,370 crore ($1.3 billion) budget approval for the IndiaAI Mission, calling it a transformative step to “catalyse India’s AI ecosystem and position it as a force shaping the future of AI for India and the world.”

The current MeitY minister, Ashwini Vaishnaw, has set an even bolder roadmap: to develop its own Graphics Processing Unit (GPU) within the next three to five years and launch a foundational AI platform within 10 months. He envisions India-made GPUs costing as low as $1 per hour, significantly lowering the cost of AI research and deployment. However, with GPU development requiring five years, the government has allocated only Rs 2,000 crore for the current fiscal year—a modest sum compared to investments by leading AI nations.

Despite financial constraints, AI adoption in India is already high. Reports show that Indian students and professionals are 30 per cent more likely to use generative AI than their Asia-Pacific peers. Over four lakh IT
professionals currently work with AI, mainly servicing clients in the US, Canada, and Australia. However, if India aims to transition from being a service provider to a global AI innovator, it must overcome critical challenges in infrastructure, funding, and research.

One of the biggest hurdles in developing a domestic AI platform is the absence of a centralised government agency overseeing AI innovation. For instance, India’s space success was driven by ISRO, a state-run agency dedicated to research and execution. However, no such institution exists for AI. While entities like MeitY, NITI Aayog, and CDAC play a role, the lack of a singular driving force slows coordination and efficiency.

A dedicated AI agency that collaborates with private IT firms could expedite the development of an indigenous AI platform by mobilising resources and streamlining investment. Given the modest annual budget allocation for the IndiaAI Mission, pooling capital from multiple stakeholders—including the government, private sector, and academia—is crucial for timely AI innovation.

Indian IT firms have long excelled at providing AI solutions as service providers to global clients. But their traditional role has focused on adapting and deploying existing technologies rather than developing foundational AI models from scratch. While over 200 startups are working on generative AI in India, the question remains: do they possess the required expertise to build a competitive AI platform within the stipulated timeframe? Even if the answer is optimistic, challenges persist. The lack of computing infrastructure, limited R&D investment, and hurdles in data acquisition are critical concerns. India must prioritise building high-performance computing clusters and AI research labs to bridge these gaps.

China’s DeepSeek AI model, launched at a fraction of the cost of OpenAI’s ChatGPT, has demonstrated that world-class AI can be built efficiently. DeepSeek charges just $0.50 monthly for advanced features, compared to ChatGPT’s $20 subscription. Meanwhile, China and the US continue improving their platforms. Between 2010 and 2022, China and the US secured 60% and 20% of global AI patents, while India had only 0.5%.

India’s ability to replicate DeepSeek’s model depends on more than just the $1.3 billion AI Mission budget --competing with approximately $3 billion of Microsoft for new AI infrastructure; it requires a shift from a service-oriented industry to deep R&D, backed by substantial public funding and strategic global partnerships. India ranks 46 out of 188 countries in the Government AI Readiness Index 2024 published by Oxford Insights, lagging far behind the US (rank 1) and China (rank 23). The report shows India significantly lags in technology, adaptability, maturity, and infrastructure. In competing with global AI leaders, India must upgrade its research infrastructure above a vast AI talent pool. 

One advantage India holds is its ability to secure GPUs at significantly lower rates. The IndiaAI Mission’s recent GPU tender saw prices averaging Rs 115.85 per GPU hour ($1.34), considerably lower than the global average of $2.50–$3.00. However, to remain competitive in the long run, India must focus on domestic chip design and negotiate with the US for unrestricted Tier 1 hardware supply. 

India’s ambition for localised development of massive AI infrastructure is also crucial from a geopolitical perspective. India has historically prioritised data sovereignty and national security. 

Indian policymakers must strategically prioritise indigenous AI platforms and technologies with broader national security objectives like other digital infrastructure projects such as UPI and ONDC. India is home to the largest AI adopters, and boosting the Indian manufacturing sector is one of the government’s priorities. Leveraging AI innovation by creating massive AI infrastructure and the required hardware for localised production can attract global production hubs to India. 

The current AI race among and between developed and developing countries will be the ‘next big thing,’ which will determine their dominance on a global scale. It is like the race of acquiring nuclear weapons or landing on the moon. However, with its current economic and societal infrastructure, India also needs to be cautious, as it comes with its own risks, such as biases in AI models, ethical concerns, and potential job displacement. But the truth is that AI is here to stay, and so is the race. The question is—how fast can India run?

(Prashant is an assistant professor in the Department of Public Policy, Manipal Academy of Higher Education, and Alok is a senior research fellow, Centre for Economic Studies and Policy, ISEC, Bengaluru)